Anticipatory anxiety is a complex combination of a future-oriented cognitive state, negative affect, and autonomic arousal. A dual-task paradigm of anticipation of electric shocks and a motor-learning task was used to examine the changes in neural patterns of activation associated with modulation of the cognitive state in anxiety by a distracting motor task. We used positron emission tomography (PET) and O-15-water to measure regional cerebral blood how (rcbf) in 10 healthy male volunteers. A 2 x 2 factorial design(shock vs no shock) x (low vs high distraction) was used with three scans per condition. Twelve PET scans were performed on each subject, In six of these scans, subjects were given electric shocks. In all scans, subjects also simultaneously performed a motor repetition (low distraction) or learning (high distraction) task. Galvanic skin conductance (GSR), Spielberger State and Trait Anxiety Inventory (STAI), and self-report data were also collected. In comparisons between the shock and no-shock conditions, the main finding was of increased rcbf in the left insula (-38,8,8) (z = 4.85, P < 0.05 corrected) and a homologous area in the right insula at a lower threshold(z = 3.20, P = 0.001 uncorrected). Other areas activated were the right superior temporal sulcus, left fusiform, and left anterior cingulate. Using the STAI-state scores as a covariate of interest, significant correlations with rCBF were seen in the left orbitofrontal cortex, left insula, and left anterior cingulate cortex. There was no significant distraction effect as measured by the STAI, self-report, GSR response or interactional analysis of the PET data. These findings support the role of paralimbic structures as neural substrates of anticipatory anxiety. The failure to demonstrate behavioral and neurophysiological changes with the distracter task may reflect the modest increases in anxiety with the shock, the relatively simple distracter task, and small sample size. (C),1999 Academic Press.